Review ArticleContribution of Neurons and Glial Cells to Complement-MediatedSynapse Removal during Development, Aging and inAlzheimer’s Disease

Celia Luchena,1 Jone Zuazo-Ibarra,1 Elena Alberdi ,1 Carlos Matute ,1

and Estibaliz Capetillo-Zarate 1,2,3

1Achucarro Basque Center for Neuroscience, CIBERNED and Departamento de Neurociencias de la Universidad del País Vasco UPV/EHU, E-48940 Leioa, Spain2IKERBASQUE, Basque Foundation for Science, 48011 Bilbao, Spain3Weill Cornell Medical College, Cornell University, New York, NY 10065, USA

Correspondence should be addressed to Estibaliz Capetillo-Zarate; estibaliz.capetillo@ehu.eus

Received 15 June 2018; Revised 13 August 2018; Accepted 24 September 2018; Published 11 November 2018

Guest Editor: Fulvio Celsi

Copyright © 2018 Celia Luchena et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Synapse loss is an early manifestation of pathology in Alzheimer’s disease (AD) and is currently the best correlate to cognitivedecline. Microglial cells are involved in synapse pruning during development via the complement pathway. Moreover, recentevidence points towards a key role played by glial cells in synapse loss during AD. However, further contribution of glial cellsand the role of neurons to synapse pathology in AD remain not well understood. This review is aimed at comprehensivelyreporting the source and/or cellular localization in the CNS—in microglia, astrocytes, or neurons—of the triggering components(C1q, C3) of the classical complement pathway involved in synapse pruning in development, adulthood, and AD.

1. Introduction

Alzheimer’s disease (AD) is a neurodegenerative disorder,which is clinically characterized by progressive cognitivedecline finally leading to the full-blown picture of dementia[1]. AD represents 50 to 70% of all dementia cases. Yet, ithas no cure. Synaptic loss and dendritic loss have beenobserved in the hippocampus and neocortex of AD patients[2]. The decrease in cortical synaptic density is mirrored bychanges in the presynaptic marker synaptophysin and corre-lates with cognitive decline in AD patients [3]. Therefore,understanding the underlying mechanisms responsible forsynapse loss during AD is of critical importance in order toidentify new therapeutic targets.

The complement system is part of the innate immunesystem in multicellular organisms, and it is activated by threebiochemical pathways (explained in more detail in Section 2).

The classical complement pathway is activated when ligandsbind to C1q triggering C1 complex activation. C3, a centralprotein of the complement cascade, acts as downstream ofC1q in the classical complement cascade and also activatesthe alternative pathway when ligands bind directly to it.Recent publications point towards a role played bymicroglia and astrocytes in early synapse pruning duringdevelopment, presumably via the classical complementpathway [4]. They also showed that expression of C1qprotein by retinal neurons modulated by astrocytes was acrucial event for synaptic pruning [5].

In AD, complement components have been associatedwith amyloid-β (Aβ) plaques [6, 7]. It has also been reportedthat oligomeric/fibrillar Aβ and hyperphosphorylated tau(pTau) activate the complement pathway by binding to C1q[8–13]. C1q is upregulated and associated to synapses in thepresence of oligomeric Aβ [8]. Under these circumstances,

HindawiMediators of InflammationVolume 2018, Article ID 2530414, 12 pageshttps://doi.org/10.1155/2018/2530414

the classical complement pathway activates and results insynapse loss before Aβ deposition takes place [8]. C3 has beenlocalized on reactive astrocytes in human AD cases [14] andthey might contribute to synapse loss by releasing comple-ment components themselves.

The origin and contribution of complement proteins tosynapse pruning in development and AD are not well under-stood and require further investigation. The present review isaimed at addressing the role of microglia and astrocytes incomplement-mediated synaptic pruning associated withdevelopment, adulthood, aging, and AD.

2. Complement-Mediated SynapsePruning during Development

During development, active synapses mature, while lessactive ones are engulfed and removed by microglia [15–18].Synapse removal is regulated, among other mechanisms, bythe classical complement cascade.

The complement system can be activated by three differ-ent pathways: (1) the classical pathway commonly is initiatedby antigen-antibody binding leading to the phagocytosisand/or pore formation in membrane, lysis, and cell death,(2) the alternative pathway is continuously active at lowlevels and not activated by pathogen or antibody binding. Itleads to opsonization and kills pathogens, and (3) the lectinpathways are activated by mannan-binding lectin as opposedto antibody-antigen recognition (reviewed in [19]). In thisreview, we will focus on the classical pathway. This pathwayis initiated when the antigen-antibody complex binds to theC1q protein. In addition to antigen antibody activation ofC1q, antibody-independent activation of C1q has been con-sistently described [9, 10, 20–22]. C1q protein is the recogni-tion subcomponent of the C1 complex, which is composed ofone molecule of C1q, two molecules of C1r, and two mole-cules of C1s (C1qr2s2). C1q recognizing and binding todifferent ligands trigger the cascade of events leading to thecleavage of complement component C3 into the fragmentsC3a and C3b. It is worth mentioning that the cleavageof C3 can also be directly activated by the alternative path-way. C3b is one of the primary complement opsonins andhas been shown to tag synapses targeted for elimination[15, 19, 23]. Further C3b cleavage gives rise to iC3b, whichsubsequently binds to complement receptor 3 (CR3) inmicroglia, in turn driving a response that, among others,promotes phagocytosis of cellular structures like synapses[24, 25]. Therefore, the cellular localization of C1q and C3and their receptors is crucial in understanding the mecha-nism regulating synapse removal.

In vitro and in vivo studies involving the use of mRNAexpression and immunohistochemistry techniques (Table 1)have described the localization of C1q in neurons, both insynaptic puncta [4, 5] and axons [5] during development.Also, astrocyte-secreted transforming growth factor-β(TGF-β) has been demonstrated to increase C1q expressionin neurons (Figure 1(a)) [5].

The expression of C3 protein has been shown to be cru-cial in promoting synapse phagocytosis [24]. However, thecellular source of C3 in the CNS during development remains

elusive. Immunohistochemical analysis revealed that C3 ispresent in synaptic puncta in wild-type mice, whereas C3knock-out (KO) mice were devoid of it [4, 15]. And, as statedabove, this led to proposing that C3 cleavage products (C3b/iC3b) could be tagging synapses for pruning [15].

During development, binding of C3 cleavage productsto CR3 promotes phagocytosis. CR3, also called CD11b/CD18 or mac-1, is a receptor for iC3b [26] and isuniquely expressed by microglia in the CNS [15, 27–29].Removal of CR3 results in increased number of synapsesin CR3 KO mice [15].

Altogether, these data suggest that during the develop-mental period of synaptic refinement, astrocyte-secretedTGF-β increases C1q expression in neurons. This resultsin the release of C3. The ligand is subsequently cleavedto yield C3b/iC3b fragments, which in turn bind to CR3in microglia ultimately promoting the engulfment of syn-apses (Figure 1(a)).

3. Synapse Pruning in Adulthood andNormal Aging

Although synapse remodeling is described to be persistentduring life, once overproduced synapses have been removedduring postnatal development, synapse elimination is down-regulated and remains stable (Figure 1(b)). In adulthood andunder healthy conditions, there is a reduction in the expres-sion of C1q and C3 components [24].

A number of in vitro and in vivo studies reported thatmicroglial cells are the main source of C1q in adulthoodand normal aging (Table 2) [30–32]. Importantly, Linnartzand collaborators showed that sialic acid, a cellular mem-brane component, was crucial in preventing C1q binding totarget molecules. The study demonstrated that desialylatedneuronal structures are marked by the complement andcleared by microglia, in a process involving CR3 [31]. Thismechanism has been related to neurites [31]. Whether thismechanism is also involved directly in synapse removalneeds to be investigated. Even though the main source ofC1q in adulthood is microglia, C1q has also been localizedat synaptic puncta in human and mouse tissue by immuno-histochemistry [32]. Furthermore, positive immunoreactivityof C1q has been observed in a subset of interneurons [32, 33].In a separate study, astrocyte-derived exosomes obtainedfrom human subjects showed expression of C1q [34].

Both microglia [30, 31, 35] and astrocytes [35, 36] havebeen reported as primary sources of C3 in primary cultures[30, 35]. Similar to C1q, C3 has also been detected inastrocyte-derived exosomes from human subjects [34]. Asindicated before, C3 cleaves into different products includ-ing C3a and C3b. C3b cleavage product iC3b binds to CR3[24, 25], whereas C3a binds to C3aR [14].

As described in Section 2, microglial cells are the onlysource of CR3 in the brain [37]. Linnartz et al. [31] reportedthat removal of neurites could be mediated by CR3 undersome circumstances as indicated before.

Finally, C3a binds to C3aR, a receptor present in neu-rons, microglia, and astrocytes [37–42]. C3a is an anaphyla-toxin linked to proinflammatory signaling and chemotaxis

2 Mediators of Inflammation

[25, 43]. In the CNS, C3a-C3aR binding has been reportedto mediate synaptic plasticity [14, 43] and microglialphagocytosis [14].

4. AD, Synapse Pathology, and Gliosis

The neuropathology associated with AD includes synapseloss, neuronal loss, neurofibrillary tangles (NFT), Aβ accu-mulation, and gliosis [44]. Neuritic senile plaques and NFTare the two main pathological hallmarks of AD. It is widelyknown that neuritic plaques are surrounded by dystrophicneurites and activated microglia and astrocytes. Dystrophicneurites appear early in the disease and are considered tobe aberrant axon/presynaptic terminals caused by (at leastin part) alterations in cytoskeleton and axonal flux [45–48].NFT are intraneuronal aggregates of pTau protein [49, 50],and NFT deposition correlates with progression of thedisease [51, 52].

Aβ, the main component of plaques, is the result of thecleavage of amyloid precursor protein (APP) and accumulatesboth intra- [53–56] and extracellularly [57, 58] leading to syn-aptic dysfunction, which is currently the best correlate of cogni-tive decline in patients [3]. Synaptic dysfunction is reported asan early manifestation of AD [59]. Loss of synaptophysin hasbeen shown to correlate well with Aβ accumulation [60, 61].In vitro studies in primary neuronal cultures from AD mousemodel Tg2576 showed that Aβ produced by proteolyticcleavage of mutant SwedishAPP is sufficient to induce synapsepathology [62], demonstrating that neurons have a cell-autonomous mechanism for synapse removal.

Aβ aggregation and accumulation associated with ADpathogenesis trigger an inflammatory response in affectedareas of the brain. Thus, active microglia and astrocytesare conspicuous around neuritic plaques [63, 64]. Bothmicroglia and astrocytes modify their morphology to adopta reactive morphology and undergo functional changes(reviewed in [65–69]). Chronic inflammation states, charac-terized by sustained reactive gliosis, have been shown toworsen the AD pathology (reviewed in [66]). In animalmodels of AD, activated microglia undergo a change in mor-phology, from a typical ramified structure to a more amoe-boid morphology. This has been associated with increasedproliferation [70, 71] and expression of inflammatorymarkers [70, 72]. Recently, in AD brain tissue samples,

region-specific microglial deterioration associated withdecrease in amount of microglia has been described indicatingpossible differential microglial responses between AD modelsand patients [68, 73]. Reactive astrocytic processes penetrateAβ deposit, fragmenting and isolating plaques from the sur-rounding neuropil [63]. Additionally, reactive glia derivedfrom neurological disorders have been characterized usingtranscriptomic profiling studies [68, 73]. Disease models sug-gest the existence of expression of specific genetic profiles for“disease-associated microglia” (DAM) [68, 74, 75].

Recently, activated glial cells have been linked to synapseloss in AD. Activated microglia modulate synapse loss via thecomplement pathway in an AD model [8]. Other studieshave shown that apolipoprotein E (APOE) isoforms controlfor C1q accumulation in the brain and modulate phagocyto-sis by astrocytes [76]. APOE4, the major genetic risk factorfor the late onset AD, negatively affects the rate of synapsepruning and turnover by astrocytes. C1q protein was signifi-cantly increased in the hippocampus of APOE4 knock-in(KI) mice when compared with APOE3 KI mice [76]. More-over, genome-wide association studies (GWAS) and anetwork-based integrative analysis have linked genes of theimmune system—like complement CR1—with increased riskof developing AD [77, 78].

Despite of neuron-autonomous and Aβ-induced synapseturnover, a number of studies demonstrate that glial cells arealso associated with synapse loss in AD. The complementsystem might play an important role in such a process.

5. Complement and Synapse Pathology in AD

Components of the classical complement pathway have beenassociated with senile plaques [79, 80], fibrillar Aβ, NFT, anddystrophic neurites [11, 63, 64, 81–85]. Oligomeric/fibrillarAβ and pTau bind to C1q and activate the complement clas-sical pathway [8–13].

Complement components have been studied as possiblebiomarkers for AD. Daborg et al. [86] found increased C3levels in AD patients compared with patients suffering frommild cognitive impairment but not progressing to AD. Theyalso showed elevated cerebrospinal fluid (CSF) CR1 levels inmild cognitive impairment that progressed to AD and ADpatients when these groups merged. Bonham et al. [87]showed a significant interaction between APOE4 and CSF

Table 1: Cellular localization of complement components during development.

Cell type Localization Experimental model Method Ref.

C1q

Neuron Synaptic puncta, axonsIn vitro: rat primary culture

In vivo: C1qaKO, Tgfbr2−/−, C1qa−/− mice

qRT-PCR, in situhybridization, ICC, IHC,

WB[4, 5]

C3

Unknown Synaptic puncta In vivo: C3KO mice IHC [4, 15]

CR3

Microglia Cell surface In vivo: CR3KO mice IHC [15]

qRT-PCR: semiquantitative PCR; ICC: immunocytochemistry; IHC: immunohistochemistry; WB: Western blot.

3Mediators of Inflammation

C3 on both CSF Aβ and CSF pTau. Their results also indi-cated that Aβ mediates C3 effect on pTau.

Schaffer et al. [15] demonstrated the key role played bymicroglia in developmental synaptic pruning. Since then,

glial cells and the complement pathway have become increas-ingly relevant in the study of synapse loss associated with AD.

Recently, a study demonstrated the involvement ofmicroglia in synapse pathology at early stages of AD,

TGF-�훽

C3b

C3b

C3b

C3a

Astrocyte

Microglia

Strongsynapse

Weaksynapse

iC3b

C3

CR3C1q

(a) Normal development

C1q

CR3C3

C3aRC1q

(b) Healthy brain

iC3bC3b

C3a

C3b

C3a

C1q

C3

A�훽C3aR

Reactiveastrocyte

Reactivemicroglia

CR3

(c) Alzheimer’s disease

Figure 1: Model of complement-mediated synapse elimination during development, adulthood, and Alzheimer’s disease. (a) During earlypostnatal development, synaptic pruning takes place in order to eliminate excessive or weak synapses. Astrocytes induce the expression ofC1q in neurons through TGF-β, and C1q colocalizes with synapses. The complement protein C3, which also colocalizes with synapticpuncta, is enzymatically cleaved to smaller fragments C3a and C3b. Finally, microglia engulf the synapse through the interaction of iC3b,the cleavage product of C3b, with its CR3 receptors. (b) In the healthy brain, synaptic pruning decreases with age to basal levels andcomplement protein expression is reduced. Nonetheless, microglia and astrocytes continuously survey surrounding synapses. (c) AD brainis characterized by progressive accumulation of extracellular and intracellular Aβ, gliosis, and neuroinflammation. Some studies havereported the role of microglia and complement pathway on synapse loss in AD models. Neuron-derived C1q and microglia-derived C1qare recruited to synapses and interact with Aβ. This triggers the activation of complement protein C3, expressed by both astrocytes andmicroglia. C3 is cleaved to smaller fragments such as C3b and iC3b that tag synapses and bind to CR3 on microglia. All these events leadto the removal of tagged synapses by the latter.

4 Mediators of Inflammation

preceding plaque formation [8], thus supporting the exis-tence of a mechanism described during development andalso modulating early pathological conditions during AD.

In AD brains, C1q has been associated with oligomeric/fibrillar Aβ and pTau, and described within neurons andactivated-microglia around plaques (Table 3; Figure 1(c))[9–13, 88]. It was not until recently that C1q was found asso-ciated with synapse pathology in AD. C1q expression wasincreased in a region-specific manner in one-month-oldJ20 AD mice [8]. A significant increase in C1q and PSD-95 colocalization was also observed in J20 mice and alsoin wild-type (WT) mice when injected with Aβ oligomers.Subsequently, at 3 months of age, a reduction of postsynap-tic marker PSD-95 and synaptic puncta was described inthese AD mice. In vitro and in vivo experiments also asso-ciated oligomeric Aβ and C1q with synaptic dysfunctionand in a common pathway that resulted in synapse removal[8]. These findings supported previously reported data indi-cating that oligomeric—but not monomeric—Aβ activatesthe complement cascade in AD [10, 12, 13] and showedthat C1q is one of the key players for synapse loss in earlypreplaque AD.

Both microglia and astrocytes have been shown assources of C1q in AD. Microglial C1qa upregulation wasreported using fluorescence in situ hybridization techniquesin the hippocampi of both J20 and Aβ oligomer-treatedWT mice [8]. Additionally, although human AD-culturedastrocytes showed low levels of C1q [84], astrocyte-derivedexosomes from human AD showed higher expression ofC1q compared to control cases [34]. Furthermore, astrocyticcultures from aged 5xFAD mice showed an increase of C1qexpression compared with those from control mice [89].

Such increase was also associated with Aβ plaques. Neverthe-less, the association between C1q expression and synapse lossobserved in the 5xFAD model requires further clarification.

As described for C1q, C3 has also been associated withsynaptic puncta in AD mice [8]. A significant increase ofC3 and PSD-95 colocalization was observed in the APP/PS1-AD model. Deletion of C3 in APP/PS1 mice pre-vented hippocampal synapse loss in 4-month-old APP/PS1 mice [8], supporting the key role played by the C3complement component in synapse loss in AD at earlystages of the disease.

A study has reported that exposure to Aβ activates astro-glial NFκB, resulting in the release of C3 by astrocytes. Thisin turn leads to reduced synaptic density and altered den-dritic morphology [14]. C3 cleavage products including C3aand C3b have been shown to mediate Aβ phagocytosis underpathological conditions [43, 90].

In AD, as showed in development and adulthood, CR3 isexclusively expressed by microglia. Deletion of CR3 has beenshown to be protective both in vitro and in vivo [8, 91]. Whileinjection of Aβ oligomers in WT mice resulted in synapseloss, Aβ oligomers failed to induce synapse loss in of CR3KO mice [8], supporting the role played by the complementin synapse removal during AD.

Additionally, a new phagocytic-independent role ofCR3 has been attributed to this receptor in AD models.APP AD mice with deleted CR3 expression presented lessAβ load and reduced interstitial soluble Aβ when comparedwith APP mice expressing normal CR3 levels [91]. Alto-gether, CR3 mediates phagocytosis-dependent Aβ plaqueand synapse removal and also phagocytosis-independentinterstitial soluble Aβ elimination. Further studies in AD

Table 2: Cellular localization of complement components in adulthood and aging.

Cell type Localization Experimental model Method Ref.

C1q

Neuron Synaptic punctaIn vivo: C57BL/6 mice

Human tissueIHC [32]

Microglia CellularIn vitro: mice primary culture

In vivo: C57BL/6, C1qa deletion miceqRT-PCR, IF, IHC, WB [30–33]

Astrocyte Exosomes Human plasma Immunoassay [34]

C3

Microglia Cellular In vitro: mice primary cultures qRT-PCR, IF, ICC, WB [30, 31, 35]

AstrocyteCellular In vitro: mice primary cultures qPCR, IF, WB, RNA-seq [30, 35, 36]

Exosomes Human plasma Immunoassay [34]

CR3

Microglia — In vitro: mice primary culture qRT-PCR, IF [31]

C3aR

Neuron CellularIn vitro: cultured neural stem cells, rat

In vivo: WT mice and ratIn situ hybridization, IF [41, 42]

Microglia CellularIn vivo: WT miceHuman tissue

In situ hybridization, IF, IMC [37–39]

Astrocyte Cellular In vitro: CB193 cell line, human astrocyte cultures IF, RT-PCR, IP [39, 40]

qRT-PCR: semiquantitative PCR; ICC: immunocytochemistry; IHC: immunohistochemistry; WB: Western blot; IF: immunofluorescence; FACS: fluorescence-activated cell sorting; IP: immunoprecipitation.

5Mediators of Inflammation

models are required to further expand the key role played byCR3 in AD progression.

Finally, not only CR3 but also microglial C3aR receptorsmight be involved in synaptic loss in AD [14, 37]. Aβ inducesastrocytic release of C3 via NFκB activation, which in turninteracts with microglial C3aR to mediate pathology in AD[14, 37]. C3aR is also expressed by neurons in AD [14].

Hence, in summary, the presence of oligomeric/fibrillarAβ at the synaptic area increases the levels of C1q and bindsto it. As a result of which, the levels of C3 increase and itscleavage product iC3b binds to microglial CR3, which in turnresults in synapse removal (Figure 1(c)).

6. CR1 and AD

CR1 is a transmembrane glycoprotein that can be found inthe plasma membrane of erythrocyte and monocyte/macro-phage among other cells and is involved in the phagocytosisof complement-opsonized pathogens [92].

In the periphery, CR1 is mainly expressed by erythrocytesand is involved in the clearance of complement-opsonizedpathogens. It has been reported that peripheral Aβ is opson-ized by the complement and captured by erythrocyte andmonocyte/macrophage CR1 [21, 93]. In AD patients, captureof Aβ by erythrocytes is reduced [21, 93] and Aβ immuno-therapy improves this clearance mechanism in vitro and inliving primates [21].

Several laboratories have studied the cellular localizationof CR1 within the brain. In 1996, Gasque and Morgan [38]reported that CR1 is expressed by primary human astrocytesand T193 astrocytes in vitro. Those results were later con-firmed using immunohistochemistry in human tissue sec-tions from normal and multiple sclerosis (MS) brains [38].However, in 1999, Singhrao et al. were unable to detectCR1 in neurons, astrocytes, or microglia [94].

In 2009, a GWAS study identified single-nucleotide poly-morphisms (SNP) in CR1 that were associated with late onsetAD and therefore as a risk factor for the disease [77]. Furtherwork confirmed this association [95–99], and the study of itscellular localization in the CNS intensified.

While Allen et al. and Karch et al. [100, 101] reportedCR1 mRNA expression on cortical homogenates of ADbrains, Holton et al. [102] found CR1 to be associated specif-ically with the frontal cortex white matter and cerebellum inAD samples. At cellular level, Hazrati et al. [103] showedCR1 immunoreactivity in neurons, choroid plexus, andblood cells of human origin, whereas CR1 localization inhuman microglia was not evident.

Based on the genetic association, expression, and func-tion, in 2016, Fonseca et al. [104] analyzed the role of CR1in AD. They showed the specificity of two monoclonal anti-CR1 antibodies for astrocytes both in human tissue samplesand in human brain-derived astrocyte cultures. Nonspecificimmunoreactivity in neurons or microglia was detected.

Table 3: Cellular localization of complement components in Alzheimer’s disease.

Cell type Localization Experimental model Method Ref.

C1q

NeuronCellular Human tissue IHC, in situ hybridization [88]

Synaptic puncta In vivo: J20 mice IF [8]

Microglia — In vivo: J20 mice IF [8]

AstrocyteCellular

In vitro: human primary cultureIn vivo: 5xFAD mice

ELISA, IF [84, 89]

Exosomes Human plasma samples Immunoassay [34]

C3

Neuron Synaptic puncta In vivo: J20 mice IF [8]

Microglia —In vitro: mice primary cultureIn vivo: APPswe/PS1dE9 mice

qPCR, IP, in situ hybridization [35, 37]

Astrocyte

—In vitro: mice primary culture

In vivo: APP/TTA, APPswe/PS1dE9 miceqPCR, in situ hybridization, IC [30, 37]

Cellular In vivo: APP mice IF [124]

— Human tissue qPCR, in situ hybridization [14]

Exosomes Human plasma Immunoassay [34]

CR3

Microglia CellularIn vitro: mice primary culture, BV2 cell line

In vivo: J20 miceIF, WB [8, 91]

C3aR

Neuron —In vitro: mice primary culture

In vivo: APP/TTA miceIF [118]

Microglia —In vitro: mice primary cultureIn vivo: APPswe/PS1dE9 mice

qPCR, in situ hybridization [37]

qRT-PCR: semiquantitative PCR; ICC: immunocytochemistry; IHC: immunohistochemistry; WB: Western blot; IF: immunofluorescence; FACS: fluorescence-activated cell sorting; IP: immunoprecipitation.

6 Mediators of Inflammation

The authors determined that the use of different antibodieswith different reactivities or levels of detection could explaindifferences among different laboratories [104]. They alsoreported that neither CR1 distribution in the brain nor itsbinding activity correlated with AD-related CR1 polymor-phisms and diagnostics. They concluded that further func-tional studies on peripheral CR1 on red cells might bringlight on how this receptor contributes to AD.

A recent study by Johansson et al. [105] investigated theperipheral interaction of CR1 with Aβ in AD and was alsounable to detect CR1 expression in the brain. By contrastand in line with previous results by their lab [93] as well asthose by Fonseca et al. [104], they showed that CR1 in eryth-rocytes was significantly reduced in AD. CR1-mediatederythrocyte capture of circulating Aβ was also significantlyreduced. Moreover, the SNP of CR1 that increases AD riskwas associated with decreased CR1 in erythrocytes, whilethe SNP of CR1 that decreases AD risk was associated withincreased CR1 in erythrocytes.

In summary, while CR1 localization in the peripheralnervous system has been strongly supported, the CR1 expres-sion in the brain is still open to debate. If it is indeedexpressed in the brain, CR1 might play a role in synapsepruning, as it is a receptor for opsonized complexes.

7. Beneficial versus Detrimental Role ofComplement in AD

Reported data indicate that the complement system mighthave a dual role in AD. Some authors argue that the comple-ment is neuroprotective. As a matter of fact, some comple-ment factors like C3 have been reported to decrease theneuropathology in hAPP mice [43]. C3 and CR3 are involvedin fibrillar Aβ phagocytosis in vitro [106]. Also, a studycarried out using APP/C3−/− mice suggested that C3 mayhave a beneficial role in plaque clearance and neuronal healthin AD. It should be noted that such study was done using 12-and 17-month-old mice [90]. C1q has also been proven tohave a neuroprotective role against Aβ toxicity both in vitroand in mice [107].

However, numerous researchers claim that the comple-ment plays a detrimental role in AD. An APP C1q KOmousemodel showed less glial activation and decreased synapse lossand neuronal degeneration [108]. Also, treatment with aC3aR antagonist that blocked the C3 signal in APP mice res-cued cognitive impairment [14]. In another study, oligomericAβ failed to induce synaptic loss in C1qa KO mice, C3 KOmice, and in mice with microglia-lacking CR3 [8]. Also, inhi-biting microglial CR3 reduced Aβ pathology [37]. Recentpublications demonstrated that lifelong C3 deficiency par-tially protects against age- and Aβ-associated hippocampalsynapse loss in mice [109, 110]. Although C3 deficiencyresulted in less Aβ clearance and, therefore, more plaques,the mice had more synapses and better cognitive function,suggesting that the immune response to Aβ was the causeof neurodegeneration. Moreover, microglia-lacking CR3 ismore efficient in degrading Aβ [91]. It has also been shownthat activation of the complement system worsens taupathology in AD models (reviewed in [111]).

All these contradictory results may be due in part tothe variety of experimental models used and the differentages at which mice were studied. Also, different brainregions were analyzed. Studies that found C3 to be benefi-cial for Aβ clearance and neuronal death were carried outusing APP [90] and hAPP [43] mice older than 10 monthswith well-established Aβ plaque pathology. However, someof the studies that found C1q, C3, and CR3 to be detri-mental for Aβ clearance and synapse health used J20 [8]and APP [91] mice at preplaque stages.

Acute versus chronic complement activation may also bea key for understanding its dual role in AD pathology. Lianet al. [14] showed that short-term treatment of microgliawith C3 or C3a promoted phagocytosis, whereas longer treat-ment diminished it.

As in AD, complement inhibition is protective in otherneurodegenerative diseases. Increased expression of C1q bymicroglia [112] and in synapses has been reported in murinemodels of glaucoma, and its inhibition prevented synapsedegeneration [113]. Moreover, C1qa was upregulated inmicroglia and neurons during West Nile virus infection andlack of C3 or C3aR in KO mice protected against virus-induced synaptic loss [114]. It has also been reported thatcomplement contributes to cell and myelin damage in MS,with an upregulation of C1q, C3, and C3-activated productsin patients’ hippocampi [115] and plaques [116].

Finally, neuroglia express inhibitory protective moleculesto prevent uncontrolled complement-mediated damage, suchas CD59, complement factor H (CFH), and complementreceptor-related protein-y (Crry), which mainly interferewith C3 [19, 117]. Modification of the expression of thesemolecules in the context of AD needs additional investiga-tion. Whether the complement plays a beneficial or a detri-mental role in AD needs further clarification.

8. Other Pathways Involved in Synapse Removal

Besides complement-mediated synapse removal, other path-ways have been described to be involved in glia-neuron inter-action that might also result in synapse elimination (reviewedin [118]). Astrocytes have been described to be implicated inactivity-dependent synapse elimination via multiple EGF-like domains 10 (MEGF10) and Mer tyrosine kinase(MERTK) pathways [119]. Recently, phagocytic capacity ofreactive astrocytes via ABCA1 and its pathway molecules,MEGF10 and GLUP1, after transient brain ischemia has alsobeen reported [120]. Also, the phagocytic clearance capacityof plaque-associated reactive astrocytes of presynapticdystrophies in an APP/PS1 AD mouse model has beendemonstrated [121].

Microglial Cx3cr1 receptor might also regulate synapticspine maintenance as reported by Paolicelli et al. [122].Cx3cr1 KO mice showed increased levels of PSD-95 andsynaptic puncta when compared to WT mice, suggesting adeficit in synaptic pruning. Also, only spontaneous vesiclerelease was observed. This was an indication for immatureconnectivity in the KO animals [122].

Additionally, a recent report [123] indicates that previ-ously described phagocytosis of synapses by microglia might

7Mediators of Inflammation

be only partial and involves specifically presynaptic termi-nals. The role of complement cascade in this selective phago-cytosis and the requirement of glial cells need to beinvestigated. Additional research including knowledge aboutthe involved signaling pathways as well as the effect of genet-ics, aging, disease, and brain region among others is neededto understand glia-mediated synaptic elimination.

9. Conclusions

Understanding the biology behind synapse loss in AD iscrucial to the discovery of new and more efficient therapeu-tic targets, ultimately aiming to stop or reverse early stagesof disease progression. In vitro studies showed that bothintracellular [53–56] and extracellular [57, 58] Aβ are associ-ated to synapse pathology. As a matter of fact, Aβ accumula-tion itself is capable of inducing synapse loss in isolatedneurons. However, in vivo studies have demonstrated theimportant contribution of both microglia and astrocytes toAD-related synapse loss. Recent studies point towards a keyrole played by the complement pathway, specifically by C1qand/or C3 initiator proteins, as the mechanism by which glialcells modulate synapse pruning. Such mechanism may takeplace not only during development but also during AD-related neurodegeneration. Therefore, different effectorsmight be contributing to AD-associated synapse loss viadifferent mechanisms.

The contribution of neurons and glial cells to synapseremoval, both in health and during neurodegeneration,remains not well understood and needs further investigation.A number of studies suggest that there might be additionalfactors involved in the regulation of synaptic pruning. Thus,new knowledge on cellular and molecular contributors willshed light to the understanding of this complex machinery.

Additional Points

Highlights. Amyloid-induced synapse loss in AD may occurvia both neuron-autonomous mechanisms and nonautono-mous mechanisms involving glial cells. Glial cells contributechiefly to synapse pruning via the complement pathway.The complement pathway plays a key role in glia-mediatedsynapse pruning. Glial contribution as a source of comple-ment components differs during development, adulthood,normal aging, and Alzheimer’s disease.

Conflicts of Interest

The authors declare that there is no conflict of interestregarding the publication of this paper.

Authors’ Contributions

Celia Luchena and Jone Zuazo-Ibarra contributed equally tothis work.

Acknowledgments

The authors thank Dr. Baleriola at Achucarro BasqueCenter for Neuroscience (Bilbao, Spain) and Dr. Solé-Domènech and Dr. Pipalia at Weill Cornell Medical College(Cornell University, New York, USA) for the helpful and crit-ical revision of the manuscript. This study was supported byCIBERNED and by grants from Ministerio de Economía yCompetitividad (SAF2016–75292-R), Gobierno Vasco(PIBA PI-2016-1-009-0016 and ELKARTEK 2016-00033),Ikerbasque, Basque Foundation for Science, and Universidaddel País Vasco/Euskal Herriko Unibertsitatea UPV/EHU.Jone Zuazo held a fellowship from Gobierno Vasco and CeliaLuchena from Fundación Tatiana Pérez de Guzmán el Bueno.

References

[1] C. Duyckaerts and D. Dickson, “Neuropathology of Alzhei-mer’s disease,” in Neurodegener, D. Dickson, Ed., pp. 47–65,Mol. Pathol. Dement. Mov. Disord. ISN Neuropath Press,Basel, 2003.

[2] I. Catalá, I. Ferrer, E. Galofré, and I. Fábregues, “Decreasednumbers of dendritic spines on cortical pyramidal neuronsin dementia. A quantitative Golgi study on biopsy samples,”Human Neurobiology, vol. 6, no. 4, pp. 255–259, 1988.

[3] S. T. DeKosky, S. W. Scheff, and S. D. Styren, “Structuralcorrelates of cognition in dementia: quantification andassessment of synapse change,” Neurodegeneration, vol. 5,no. 4, pp. 417–421, 1996.

[4] B. Stevens, N. J. Allen, L. E. Vazquez et al., “The classical com-plement cascade mediates CNS synapse elimination,” Cell,vol. 131, no. 6, pp. 1164–1178, 2007.

[5] A. R. Bialas and B. Stevens, “TGF-β signaling regulates neuro-nal C1q expression and developmental synaptic refinement,”Nature Neuroscience, vol. 16, no. 12, pp. 1773–1782, 2013.

[6] P. Eikelenboom, C. E. Hack, W. Kamphorst, and J. M.Rozemuller, “Distribution pattern and functional state ofcomplement proteins and alpha 1-antichymotrypsin incerebral beta/A4 deposits in Alzheimer’s disease,” Researchin Immunology, vol. 143, no. 6, pp. 617–620, 1992.

[7] S. E. Stoltzner, T. J. Grenfell, C. Mori et al., “Temporal accrualof complement proteins in amyloid plaques in Down’s syn-drome with Alzheimer’s disease,” The American Journal ofPathology, vol. 156, no. 2, pp. 489–499, 2000.

[8] S. Hong, V. F. Beja-Glasser, B. M. Nfonoyim et al., “Comple-ment and microglia mediate early synapse loss in Alzheimermouse models,” Science, vol. 352, no. 6286, pp. 712–716,2016.

[9] H. Jiang, D. Burdick, C. G. Glabe, C. W. Cotman, and A. J.Tenner, “Beta-amyloid activates complement by binding toa specific region of the collagen-like domain of the C1q Achain,” Journal of Immunology, vol. 152, no. 10, pp. 5050–5059, 1994.

[10] J. Rogers, N. R. Cooper, S. Webster et al., “Complement acti-vation by beta-amyloid in Alzheimer disease,” Proceedings ofthe National Academy of Sciences, vol. 89, no. 21, pp. 10016–10020, 1992.

[11] Y. Shen, L.-F. Lue, L.-B. Yang et al., “Complement activationby neurofibrillary tangles in Alzheimer’s disease,” Neurosci-ence Letters, vol. 305, no. 3, pp. 165–168, 2001.

8 Mediators of Inflammation

[12] S. W. Snyder, G. T. Wang, L. Barrett et al., “Complement C1qdoes not bind monomeric beta-amyloid,” ExperimentalNeurology, vol. 128, no. 1, pp. 136–142, 1994.

[13] S. Webster, C. Glabe, and J. Rogers, “Multivalent binding ofcomplement protein C1q to the amyloid β-peptide (Aβ) pro-motes the nucleation phase of Aβ aggregation,” Biochemicaland Biophysical Research Communications, vol. 217, no. 3,pp. 869–875, 1995.

[14] H. Lian, L. Yang, A. Cole et al., “NFκB-activated astroglialrelease of complement C3 compromises neuronal mor-phology and function associated with Alzheimer’s disease,”Neuron, vol. 85, no. 1, pp. 101–115, 2015.

[15] D. P. Schafer, E. K. Lehrman, A. G. Kautzman et al., “Microg-lia sculpt postnatal neural circuits in an activity andcomplement-dependent manner,” Neuron, vol. 74, no. 4,pp. 691–705, 2012.

[16] A. Sierra, M.-Ã. Tremblay, and H. Wake, “Never-restingmicroglia: physiological roles in the healthy brain and patho-logical implications,” Frontiers in Cellular Neuroscience,vol. 8, 2014.

[17] M.-È. Tremblay, R. L. Lowery, and A. K. Majewska, “Micro-glial interactions with synapses are modulated by visual expe-rience,” PLoS Biology, vol. 8, no. 11, article e1000527, 2010.

[18] H. Wake, A. J. Moorhouse, S. Jinno, S. Kohsaka, andJ. Nabekura, “Resting microglia directly monitor the func-tional state of synapses in vivo and determine the fate ofischemic terminals,” The Journal of Neuroscience, vol. 29,no. 13, pp. 3974–3980, 2009.

[19] A. H. Stephan, B. A. Barres, and B. Stevens, “The complementsystem: an unexpected role in synaptic pruning during devel-opment and disease,” Annual Review of Neuroscience, vol. 35,no. 1, pp. 369–389, 2012.

[20] B. M. Bradt, W. P. Kolb, and N. R. Cooper, “Complement-dependent proinflammatory properties of the Alzheimer’sdisease beta-peptide,” The Journal of Experimental Medicine,vol. 188, no. 3, pp. 431–438, 1998.

[21] W. D. Brubaker, A. Crane, J. U. Johansson et al., “Peripheralcomplement interactions with amyloid β peptide: erythrocyteclearance mechanisms,” Alzheimers Dement, vol. 13, no. 12,pp. 1397–1409, 2017.

[22] N. R. Cooper, F. C. Jensen, R. M. Welsh Jr, and M. B.Oldstone, “Lysis of RNA tumor viruses by human serum:direct antibody-independent triggering of the classicalcomplement pathway,” The Journal of Experimental Medi-cine, vol. 144, no. 4, pp. 970–984, 1976.

[23] J. J. Alexander, A. J. Anderson, S. R. Barnum, B. Stevens, andA. J. Tenner, “The complement cascade: Yin-Yang in neuro-inflammation - neuro-protection and -degeneration,” Journalof Neurochemistry, vol. 107, no. 5, pp. 1169–1187, 2008.

[24] J. Presumey, A. R. Bialas, and M. C. Carroll, “Complementsystem in neural synapse elimination in development anddisease,” Advances in Immunology, vol. 135, pp. 53–79, 2017.

[25] D. Ricklin, G. Hajishengallis, K. Yang, and J. D. Lambris,“Complement: a key system for immune surveillance andhomeostasis,” Nature Immunology, vol. 11, no. 9, pp. 785–797, 2010.

[26] D. R. Mathern and P. S. Heeger, “Molecules great and small:the complement system,” Clinical Journal of the AmericanSociety of Nephrology, vol. 10, no. 9, pp. 1636–1650, 2015.

[27] M. B. Graeber, “Changing face of microglia,” Science, vol. 330,no. 6005, pp. 783–788, 2010.

[28] B. Linnartz and H. Neumann, “Microglial activatory (immu-noreceptor tyrosine-based activation motif)- and inhibitory(immunoreceptor tyrosine-based inhibition motif)-signalingreceptors for recognition of the neuronal glycocalyx,” Glia,vol. 61, no. 1, pp. 37–46, 2013.

[29] R. M. Ransohoff and V. H. Perry, “Microglial physiology:unique stimuli, specialized responses,” Annual Review ofImmunology, vol. 27, no. 1, pp. 119–145, 2009.

[30] S. Haga, T. Aizawa, T. Ishii, and K. Ikeda, “Complement geneexpression in mouse microglia and astrocytes in culture:comparisons with mouse peritoneal macrophages,” Neurosci-ence Letters, vol. 216, no. 3, pp. 191–194, 1996.

[31] B. Linnartz, J. Kopatz, A. J. Tenner, and H. Neumann, “Sialicacid on the neuronal glycocalyx prevents complement C1binding and complement receptor-3-mediated removal bymicroglia,” The Journal of Neuroscience, vol. 32, no. 3,pp. 946–952, 2012.

[32] A. H. Stephan, D. V. Madison, J. M. Mateos et al., “A dra-matic increase of C1q protein in the CNS during normalaging,” The Journal of Neuroscience, vol. 33, no. 33,pp. 13460–13474, 2013.

[33] M. I. Fonseca, S.-H. Chu, M. X. Hernandez et al., “Cell-spe-cific deletion of C1qa identifies microglia as the dominantsource of C1q in mouse brain,” Journal of Neuroinflamma-tion, vol. 14, no. 1, article 48, 2017.

[34] E. J. Goetzl, J. B. Schwartz, E. L. Abner, G. A. Jicha, andD. Kapogiannis, “High complement levels in astrocyte-derived exosomes of Alzheimer disease,” Annals of Neurol-ogy, vol. 83, no. 3, pp. 544–552, 2018.

[35] S. Haga, K. Ikeda, M. Sato, and T. Ishii, “Synthetic Alzheimeramyloid β/A4 peptides enhance production of complementC3 component by cultured microglial cells,” Brain Research,vol. 601, no. 1-2, pp. 88–94, 1993.

[36] M. M. Boisvert, G. A. Erikson, M. N. Shokhirev, and N. J.Allen, “The aging astrocyte transcriptome from multipleregions of the mouse brain,” Cell Reports, vol. 22, no. 1,pp. 269–285, 2018.

[37] H. Lian, A. Litvinchuk, A. C.-A. Chiang, N. Aithmitti, J. L.Jankowsky, and H. Zheng, “Astrocyte-microglia cross talkthrough complement activation modulates amyloid pathol-ogy in mouse models of Alzheimer’s disease,” The Journalof Neuroscience, vol. 36, no. 2, pp. 577–589, 2016.

[38] P. Gasque and B. P. Morgan, “Complement regulatoryprotein expression by a human oligodendrocyte cell line:cytokine regulation and comparison with astrocytes,”Immunology, vol. 89, no. 3, pp. 338–347, 1996.

[39] P. Gasque, S. K. Singhrao, J. W. Neal et al., “The receptor forcomplement anaphylatoxin C3a is expressed by myeloid cellsand nonmyeloid cells in inflamed human central nervoussystem: analysis in multiple sclerosis and bacterial meningi-tis,” The Journal of Immunology, vol. 160, no. 7, pp. 3543–3554, 1998.

[40] A. Ischenko, S. Sayah, C. Patte et al., “Expression of afunctional anaphylatoxin C3a receptor by astrocytes,” Jour-nal of Neurochemistry, vol. 71, no. 6, pp. 2487–2496, 1998.

[41] S. Nataf, P. F. Stahel, N. Davoust, and S. R. Barnum, “Com-plement anaphylatoxin receptors on neurons: new tricks forold receptors?,” Trends in Neurosciences, vol. 22, no. 9,pp. 397–402, 1999.

[42] Y. Rahpeymai, M. A. Hietala, U. Wilhelmsson et al., “Com-plement: a novel factor in basal and ischemia-induced

9Mediators of Inflammation

neurogenesis,” The EMBO Journal, vol. 25, no. 6, pp. 1364–1374, 2006.

[43] T. Wyss-Coray, F. Yan, A. H.-T. Lin et al., “Prominentneurodegeneration and increased plaque formation incomplement-inhibited Alzheimer’s mice,” Proceedings ofthe National Academy of Sciences, vol. 99, no. 16,pp. 10837–10842, 2002.

[44] J. Hardy and D. J. Selkoe, “The amyloid hypothesis ofAlzheimer’s disease: progress and problems on the roadto therapeutics,” Science, vol. 297, no. 5580, pp. 353–356, 2002.

[45] T. C. Dickson, C. E. King, G. H. McCormack, and J. C. Vick-ers, “Neurochemical diversity of dystrophic neurites in theearly and late stages of Alzheimer’s disease,” ExperimentalNeurology, vol. 156, no. 1, pp. 100–110, 1999.

[46] K. R. Sadleir, P. C. Kandalepas, V. Buggia-Prévot, D. A. Nich-olson, G. Thinakaran, and R. Vassar, “Presynaptic dystrophicneurites surrounding amyloid plaques are sites of microtu-bule disruption, BACE1 elevation, and increased Aβ genera-tion in Alzheimer’s disease,” Acta Neuropathologica,vol. 132, no. 2, pp. 235–256, 2016.

[47] R. Sanchez-Varo, L. Trujillo-Estrada, E. Sanchez-Mejias et al.,“Abnormal accumulation of autophagic vesicles correlateswith axonal and synaptic pathology in young Alzheimer’smice hippocampus,” Acta Neuropathologica, vol. 123, no. 1,pp. 53–70, 2012.

[48] J. H. Su, B. J. Cummings, and C. W. Cotman, “Plaque biogen-esis in brain aging and Alzheimer’s disease. II. Progressivetransformation and developmental sequence of dystrophicneurites,” Acta Neuropathologica, vol. 96, no. 5, pp. 463–471, 1998.

[49] I. Grundke-Iqbal, K. Iqbal, M. Quinlan, Y. C. Tung, M. S.Zaidi, and H. M. Wisniewski, “Microtubule-associated pro-tein tau. A component of Alzheimer paired helical filaments,”Journal of Biological Chemistry, vol. 261, no. 13, pp. 6084–6089, 1986.

[50] K. Iqbal, I. Grundke-Iqbal, A. J. Smith, L. George, Y. C. Tung,and T. Zaidi, “Identification and localization of a tau peptideto paired helical filaments of Alzheimer disease,” Proceedingsof the National Academy of Sciences of the United States ofAmerica, vol. 86, no. 14, pp. 5646–5650, 1989.

[51] I. Alafuzoff, K. Iqbal, H. Friden, R. Adolfsson, andB. Winblad, “Histopathological criteria for progressivedementia disorders: clinical-pathological correlation andclassification by multivariate data analysis,” Acta Neuro-pathologica, vol. 74, no. 3, pp. 209–225, 1987.

[52] P. V. Arriagada, J. H. Growdon, E. T. Hedley-Whyte, andB. T. Hyman, “Neurofibrillary tangles but not senile pla-ques parallel duration and severity of Alzheimer’s disease,”Neurology, vol. 42, no. 3, pp. 631–631, 1992.

[53] E. Capetillo-Zarate, L. Gracia, F. Yu et al., “High-resolution3D reconstruction reveals intra-synaptic amyloid fibrils,”The American Journal of Pathology, vol. 179, no. 5,pp. 2551–2558, 2011.

[54] G. K. Gouras, D. Tampellini, R. H. Takahashi, andE. Capetillo-Zarate, “Intraneuronal beta-amyloid accumula-tion and synapse pathology in Alzheimer’s disease,” ActaNeuropathologica, vol. 119, no. 5, pp. 523–541, 2010.

[55] R. H. Takahashi, E. E. Nam, M. Edgar, and G. K. Gouras,“Alzheimer beta-amyloid peptides: normal and abnormallocalization,” Histology and Histopathology, vol. 17, no. 1,pp. 239–246, 2002.

[56] D. Tampellini, N. Rahman, M. T. Lin, E. Capetillo-Zarate,and G. K. Gouras, “Impaired -amyloid secretion in Alzhei-mer’s disease pathogenesis,” The Journal of Neuroscience,vol. 31, no. 43, pp. 15384–15390, 2011.

[57] A. Deshpande, H. Kawai, R. Metherate, C. G. Glabe, andJ. Busciglio, “A role for synaptic zinc in activity-dependentAbeta oligomer formation and accumulation at excitatorysynapses,” The Journal of Neuroscience, vol. 29, no. 13,pp. 4004–4015, 2009.

[58] P. N. Lacor, M. C. Buniel, L. Chang et al., “Synaptic targetingby Alzheimer’s-related amyloid beta oligomers,” The Journalof Neuroscience, vol. 24, no. 45, pp. 10191–10200, 2004.

[59] D. J. Selkoe, “Alzheimer’s disease is a synaptic failure,” Sci-ence, vol. 298, no. 5594, pp. 789–791, 2002.

[60] L. Mucke, E. Masliah, G. Q. Yu et al., “High-level neuronalexpression of Aβ1–42in wild-type human amyloid proteinprecursor transgenic mice: synaptotoxicity without plaqueformation,” The Journal of Neuroscience, vol. 20, no. 11,pp. 4050–4058, 2000.

[61] S. Oddo, A. Caccamo, J. D. Shepherd et al., “Triple-trans-genic model of Alzheimer’s disease with plaques and tan-gles,” Neuron, vol. 39, no. 3, pp. 409–421, 2003.

[62] C. G. Almeida, D. Tampellini, R. H. Takahashi et al., “Beta-amyloid accumulation in APP mutant neurons reducesPSD-95 and GluR1 in synapses,” Neurobiology of Disease,vol. 20, no. 2, pp. 187–198, 2005.

[63] S. Itagaki, P. L. McGeer, H. Akiyama, S. Zhu, and D. Selkoe,“Relationship of microglia and astrocytes to amyloid depositsof Alzheimer disease,” Journal of Neuroimmunology, vol. 24,no. 3, pp. 173–182, 1989.

[64] P. L. McGeer, S. Itagaki, H. Tago, and E. G. McGeer, “Reac-tive microglia in patients with senile dementia of the Alzhei-mer type are positive for the histocompatibility glycoproteinHLA-DR,” Neuroscience Letters, vol. 79, no. 1-2, pp. 195–200, 1987.

[65] V. Calsolaro and P. Edison, “Neuroinflammation in Alzhei-mer’s disease: current evidence and future directions,” Alz-heimer's & Dementia, vol. 12, no. 6, pp. 719–732, 2016.

[66] A. C. Cuello, “Early and late CNS inflammation in Alzhei-mer’s disease: two extremes of a continuum?,” Trends inPharmacological Sciences, vol. 38, no. 11, pp. 956–966, 2017.

[67] F. L. Heppner, R. M. Ransohoff, and B. Becher, “Immuneattack: the role of inflammation in Alzheimer disease,”Nature Reviews Neuroscience, vol. 16, no. 6, pp. 358–372,2015.

[68] V. Navarro, E. Sanchez-Mejias, S. Jimenez et al., “Microglia inAlzheimer’s disease: activated, dysfunctional or degenera-tive,” Frontiers in Aging Neuroscience, vol. 10, 2018.

[69] R. M. Ransohoff, “How neuroinflammation contributes toneurodegeneration,” Science, vol. 353, no. 6301, pp. 777–783, 2016.

[70] W. Kamphuis, L. Kooijman, S. Schetters, M. Orre, and E. M.Hol, “Transcriptional profiling of CD11c-positive microgliaaccumulating around amyloid plaques in a mouse modelfor Alzheimer’s disease,” Biochimica et Biophysica Acta(BBA) - Molecular Basis of Disease, vol. 1862, no. 10,pp. 1847–1860, 2016.

[71] A. Olmos-Alonso, S. T. T. Schetters, S. Sri et al., “Pharmaco-logical targeting of CSF1R inhibits microglial proliferationand prevents the progression of Alzheimer’s-like pathology,”Brain, vol. 139, no. 3, pp. 891–907, 2016.

10 Mediators of Inflammation

[72] E. Martin, C. Boucher, B. Fontaine, and C. Delarasse, “Dis-tinct inflammatory phenotypes of microglia and monocyte-derived macrophages in Alzheimer’s disease models: effectsof aging and amyloid pathology,” Aging Cell, vol. 16, no. 1,pp. 27–38, 2017.

[73] E. Sanchez-Mejias, V. Navarro, S. Jimenez et al., “Solublephospho-tau from Alzheimer’s disease hippocampus drivesmicroglial degeneration,” Acta Neuropathologica, vol. 132,no. 6, pp. 897–916, 2016.

[74] H. Keren-Shaul, A. Spinrad, A. Weiner et al., “A uniquemicroglia type associated with restricting development ofAlzheimer’s disease,” Cell, vol. 169, no. 7, pp. 1276–1290.e17, 2017.

[75] S. A. Liddelow, K. A. Guttenplan, L. E. Clarke et al., “Neuro-toxic reactive astrocytes are induced by activated microglia,”Nature, vol. 541, no. 7638, pp. 481–487, 2017.

[76] W.-S. Chung, P. B. Verghese, C. Chakraborty et al., “Novelallele-dependent role for APOE in controlling the rate of syn-apse pruning by astrocytes,” Proceedings of the NationalAcademy of Sciences, vol. 113, no. 36, pp. 10186–10191, 2016.

[77] J.-C. Lambert, the European Alzheimer's Disease InitiativeInvestigators, S. Heath et al., “Genome-wide associationstudy identifies variants at CLU and CR1 associated with Alz-heimer’s disease,” Nature Genetics, vol. 41, no. 10, pp. 1094–1099, 2009.

[78] B. Zhang, C. Gaiteri, L.-G. Bodea et al., “Integrated systemsapproach identifies genetic nodes and networks in late-onset Alzheimer’s disease,” Cell, vol. 153, no. 3, pp. 707–720, 2013.

[79] P. Eikelenboom, C. E. Hack, J. M. Rozemuller, and F. C. Stam,“Complement activation in amyloid plaques in Alzheimer’sdementia,” Virchows Archiv B Cell Pathology IncludingMolecular Pathology, vol. 56, no. 4, pp. 259–262, 1989.

[80] P. Eikelenboom and F. C. Stam, “Immunoglobulins and com-plement factors in senile plaques,” Acta Neuropathologica,vol. 57, no. 2-3, pp. 239–242, 1982.

[81] P. L. McGeer, H. Akiyama, S. Itagaki, and E. G. McGeer,“Activation of the classical complement pathway in brain tis-sue of Alzheimer patients,” Neuroscience Letters, vol. 107,no. 1-3, pp. 341–346, 1989.

[82] P. L. McGeer, H. Akiyama, S. Itagaki, and E. G. McGeer,“Immune system response in Alzheimer’s disease,” TheCanadian Journal of Neurological Sciences, vol. 16, no. S4,pp. 516–527, 1989.

[83] R. Veerhuis, M. J. Van Breemen, J. M. Hoozemans et al.,“Amyloid beta plaque-associated proteins C1q and SAPenhance the Abeta1-42 peptide-induced cytokine secretionby adult human microglia in vitro,” Acta Neuropathologica,vol. 105, no. 2, pp. 135–144, 2003.

[84] R. Veerhuis, I. Janssen, C. J. A. de Groot, F. L. van Muiswin-kel, C. Erik Hack, and P. Eikelenboom, “Cytokines associatedwith amyloid plaques in Alzheimer’s disease brain stimulatehuman glial and neuronal cell cultures to secrete early com-plement proteins, but not C1-inhibitor,” Experimental Neu-rology, vol. 160, no. 1, pp. 289–299, 1999.

[85] R. Veerhuis, I. Janssen, C. E. Hack, and P. Eikelenboom,“Early complement components in Alzheimer’s diseasebrains,” Acta Neuropathologica, vol. 91, no. 1, pp. 53–60,1996.

[86] J. Daborg, U. Andreasson, M. Pekna et al., “Cerebrospinalfluid levels of complement proteins C3, C4 and CR1 in

Alzheimer’s disease,” Journal of Neural Transmission,vol. 119, no. 7, pp. 789–797, 2012.

[87] for the Alzheimer’s Disease Neuroimaging Initiative, L. W.Bonham, R. S. Desikan, J. S. Yokoyama, and Alzheimer’sDisease Neuroimaging Initiative, “The relationship betweencomplement factor C3, APOE ε4, amyloid and tau in Alzhei-mer’s disease,” Acta Neuropathologica Communications,vol. 4, no. 1, article 65, 2016.

[88] A. Afagh, B. J. Cummings, D. H. Cribbs, C. W. Cotman, andA. J. Tenner, “Localization and cell association of C1q in Alz-heimer’s disease brain,” Experimental Neurology, vol. 138,no. 1, pp. 22–32, 1996.

[89] T. Iram, D. Trudler, D. Kain et al., “Astrocytes from oldAlzheimer’s disease mice are impaired in Aβ uptake andin neuroprotection,” Neurobiology of Disease, vol. 96,pp. 84–94, 2016.

[90] M. Maier, Y. Peng, L. Jiang, T. J. Seabrook, M. C. Carroll, andC. A. Lemere, “Complement C3 deficiency leads to acceler-ated amyloid beta plaque deposition and neurodegenerationand modulation of the microglia/macrophage phenotype inamyloid precursor protein transgenic mice,” The Journal ofNeuroscience, vol. 28, no. 25, pp. 6333–6341, 2008.

[91] E. Czirr, N. A. Castello, K. I. Mosher et al., “Microglial com-plement receptor 3 regulates brain Aβ levels through secretedproteolytic activity,” The Journal of Experimental Medicine,vol. 214, no. 4, pp. 1081–1092, 2017.

[92] R. Khera and N. Das, “Complement receptor 1: disease asso-ciations and therapeutic implications,” Molecular Immunol-ogy, vol. 46, no. 5, pp. 761–772, 2009.

[93] J. Rogers, R. Li, D. Mastroeni et al., “Peripheral clearance ofamyloid beta peptide by complement C3-dependent adher-ence to erythrocytes,” Neurobiology of Aging, vol. 27, no. 12,pp. 1733–1739, 2006.

[94] S. K. Singhrao, J. W. Neal, N. K. Rushmere, B. P. Morgan, andP. Gasque, “Differential expression of individual complementregulators in the brain and choroid plexus,” Laboratory Inves-tigation, vol. 79, no. 10, pp. 1247–1259, 1999.

[95] J. J. Corneveaux, A. J. Myers, A. N. Allen et al., “Associationof CR1, CLU and PICALM with Alzheimer’s disease in acohort of clinically characterized and neuropathologicallyverified individuals,” Human Molecular Genetics, vol. 19,no. 16, pp. 3295–3301, 2010.

[96] D. Harold, R. Abraham, P. Hollingworth et al., “Genome-wide association study identifies variants at CLU andPICALM associated with Alzheimer’s disease,”Nature Genet-ics, vol. 41, no. 10, pp. 1088–1093, 2009.

[97] C. Jin, W. Li, J. Yuan, W. Xu, and Z. Cheng, “Association ofthe CR1 polymorphism with late-onset Alzheimer’s diseasein Chinese Han populations: a meta-analysis,” NeuroscienceLetters, vol. 527, no. 1, pp. 46–49, 2012.

[98] B. T. Keenan, J. M. Shulman, L. B. Chibnik et al., “A codingvariant in CR1 interacts with APOE-ε4 to influence cognitivedecline,” Human Molecular Genetics, vol. 21, no. 10,pp. 2377–2388, 2012.

[99] O. Pedraza, M. Allen, K. Jennette et al., “Evaluation of mem-ory endophenotypes for association with CLU, CR1, andPICALM variants in black and white subjects,” AlzheimersDement, vol. 10, no. 2, pp. 205–213, 2014.

[100] M. Allen, M. Kachadoorian, M. M. Carrasquillo et al., “Late-onset Alzheimer disease risk variants mark brain regulatoryloci,” Neurology Genetics, vol. 1, no. 2, article e15, 2015.

11Mediators of Inflammation

[101] C. M. Karch, A. T. Jeng, P. Nowotny, J. Cady, C. Cruchaga,and A. M. Goate, “Expression of novel Alzheimer’s diseaserisk genes in control and Alzheimer’s disease brains,” PLoSOne, vol. 7, no. 11, article e50976, 2012.

[102] P. Holton, M. Ryten, M. Nalls et al., “Initial assessment of thepathogenic mechanisms of the recently identified Alzheimerrisk loci,” Annals of Human Genetics, vol. 77, no. 2, pp. 85–105, 2013.

[103] L.-N. Hazrati, C. Van Cauwenberghe, P. L. Brooks et al.,“Genetic association of CR1 with Alzheimer’s disease: a ten-tative disease mechanism,” Neurobiology of Aging, vol. 33,no. 12, pp. 2949.e5–2949.e12, 2012.

[104] M. I. Fonseca, S. Chu, A. L. Pierce et al., “Analysis of the puta-tive role of CR1 in Alzheimer’s disease: genetic association,expression and function,” PLoS One, vol. 11, no. 2, articlee0149792, 2016.

[105] J. U. Johansson, W. D. Brubaker, H. Javitz et al., “Peripheralcomplement interactions with amyloid β peptide in Alzhei-mer’s disease: polymorphisms, structure, and function ofcomplement receptor 1,” Alzheimer's & Dementia, 2018.

[106] H. Fu, B. Liu, J. L. Frost et al., “Complement component C3and complement receptor type 3 contribute to the phagocyto-sis and clearance of fibrillar Aβ by microglia,” Glia, vol. 60,no. 6, pp. 993–1003, 2012.

[107] M. E. Benoit, M. X. Hernandez, M. L. Dinh, F. Benavente,O. Vasquez, and A. J. Tenner, “C1q-induced LRP1B andGPR6 proteins expressed early in Alzheimer disease mousemodels, are essential for the C1q-mediated protection againstamyloid-β neurotoxicity,” The Journal of Biological Chemis-try, vol. 288, no. 1, pp. 654–665, 2013.

[108] M. I. Fonseca, J. Zhou, M. Botto, and A. J. Tenner, “Absenceof C1q leads to less neuropathology in transgenic mousemodels of Alzheimer’s disease,” The Journal of Neuroscience,vol. 24, no. 29, pp. 6457–6465, 2004.

[109] Q. Shi, S. Chowdhury, R. Ma et al., “Complement C3 defi-ciency protects against neurodegeneration in aged plaque-rich APP/PS1 mice,” Science Translational Medicine, vol. 9,no. 392, article eaaf6295, 2017.

[110] Q. Shi, K. J. Colodner, S. B. Matousek et al., “ComplementC3-deficient mice fail to display age-related hippocampaldecline,” The Journal of Neuroscience, vol. 35, no. 38,pp. 13029–13042, 2015.

[111] D. V. Hansen, J. E. Hanson, and M. Sheng, “Microglia inAlzheimer’s disease,” The Journal of Cell Biology,vol. 217, no. 2, pp. 459–472, 2018.

[112] G. R. Howell, D. G. Macalinao, G. L. Sousa et al., “Molecularclustering identifies complement and endothelin induction asearly events in a mouse model of glaucoma,” The Journal ofClinical Investigation, vol. 121, no. 4, pp. 1429–1444, 2011.

[113] P. A. Williams, J. R. Tribble, K. W. Pepper et al., “Inhibitionof the classical pathway of the complement cascade preventsearly dendritic and synaptic degeneration in glaucoma,”Molecular Neurodegeneration, vol. 11, no. 1, article 26, 2016.

[114] M. J. Vasek, C. Garber, D. Dorsey et al., “A complement-microglial axis drives synapse loss during virus-inducedmemory impairment,” Nature, vol. 534, no. 7608, pp. 538–543, 2016.

[115] I. Michailidou, J. G. P. Willems, E.-J. Kooi et al., “Comple-ment C1q-C3-associated synaptic changes in multiple sclero-sis hippocampus,” Annals of Neurology, vol. 77, no. 6,pp. 1007–1026, 2015.

[116] G. Ingram, S. Loveless, O. W. Howell et al., “Complementactivation in multiple sclerosis plaques: an immunohisto-chemical analysis,” Acta Neuropathologica Communications,vol. 2, no. 1, article 53, 2014.

[117] A. P. Sjöberg, L. A. Trouw, and A. M. Blom, “Complementactivation and inhibition: a delicate balance,” Trends inImmunology, vol. 30, no. 2, pp. 83–90, 2009.

[118] H. Lian and H. Zheng, “Signaling pathways regulatingneuron-glia interaction and their implications in Alzheimer’sdisease,” Journal of Neurochemistry, vol. 136, no. 3, pp. 475–491, 2016.

[119] W.-S. Chung, L. E. Clarke, G. X. Wang et al., “Astrocytesmediate synapse elimination through MEGF10 and MERTKpathways,” Nature, vol. 504, no. 7480, pp. 394–400, 2013.

[120] Y. M. Morizawa, Y. Hirayama, N. Ohno et al., “Reactiveastrocytes function as phagocytes after brain ischemia viaABCA1-mediated pathway,” Nature Communications,vol. 8, no. 1, article 28, 2017.

[121] A. Gomez-Arboledas, J. C. Davila, E. Sanchez-Mejias et al.,“Phagocytic clearance of presynaptic dystrophies by reactiveastrocytes in Alzheimer’s disease,” Glia, vol. 66, no. 3,pp. 637–653, 2018.

[122] R. C. Paolicelli, G. Bolasco, F. Pagani et al., “Synaptic pruningby microglia is necessary for normal brain development,” Sci-ence, vol. 333, no. 6048, pp. 1456–1458, 2011.

[123] L. Weinhard, G. di Bartolomei, G. Bolasco et al., “Microgliaremodel synapses by presynaptic trogocytosis and spine headfilopodia induction,” Nature Communications, vol. 9, no. 1,article 1228, 2018.

[124] J. Zhou, M. I. Fonseca, K. Pisalyaput, and A. J. Tenner, “Com-plement C3 and C4 expression in C1q sufficient and deficientmouse models of Alzheimer’s disease,” Journal of Neuro-chemistry, vol. 106, no. 5, pp. 2080–2092, 2008.

12 Mediators of Inflammation

Stem Cells International

Hindawiwww.hindawi.com Volume 2018

Hindawiwww.hindawi.com Volume 2018

MEDIATORSINFLAMMATION

of

EndocrinologyInternational Journal of

Hindawiwww.hindawi.com Volume 2018

Hindawiwww.hindawi.com Volume 2018

Disease Markers

Hindawiwww.hindawi.com Volume 2018

BioMed Research International

OncologyJournal of

Hindawiwww.hindawi.com Volume 2013

Hindawiwww.hindawi.com Volume 2018

Oxidative Medicine and Cellular Longevity

Hindawiwww.hindawi.com Volume 2018

PPAR Research

Hindawi Publishing Corporation http://www.hindawi.com Volume 2013Hindawiwww.hindawi.com

The Scientific World Journal

Volume 2018

Immunology ResearchHindawiwww.hindawi.com Volume 2018

Journal of

ObesityJournal of

Hindawiwww.hindawi.com Volume 2018

Hindawiwww.hindawi.com Volume 2018

Computational and Mathematical Methods in Medicine

Hindawiwww.hindawi.com Volume 2018

Behavioural Neurology

OphthalmologyJournal of

Hindawiwww.hindawi.com Volume 2018

Diabetes ResearchJournal of

Hindawiwww.hindawi.com Volume 2018

Hindawiwww.hindawi.com Volume 2018

Research and TreatmentAIDS

Hindawiwww.hindawi.com Volume 2018

Gastroenterology Research and Practice

Hindawiwww.hindawi.com Volume 2018

Parkinson’s Disease

Evidence-Based Complementary andAlternative Medicine

Volume 2018Hindawiwww.hindawi.com

Submit your manuscripts atwww.hindawi.com